[PATCH bpf] bpf: Fix off-by-one error in bpf_mem_cache_idx()

[PATCH bpf] bpf: Fix off-by-one error in bpf_mem_cache_idx()

[Hou Tao]

From: Hou Tao <houtao1@huawei.com>

According to the definition of sizes[NUM_CACHES], when the size passed
to bpf_mem_cache_size() is 256, it should return 6 instead 7.

Fixes: 7c8199e24fa0 ("bpf: Introduce any context BPF specific memory allocator.")
Signed-off-by: Hou Tao <houtao1@huawei.com>
---
 kernel/bpf/memalloc.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/kernel/bpf/memalloc.c b/kernel/bpf/memalloc.c
index ebcc3dd0fa19..1db156405b68 100644
--- a/kernel/bpf/memalloc.c
+++ b/kernel/bpf/memalloc.c
@@ -71,7 +71,7 @@ static int bpf_mem_cache_idx(size_t size)
 	if (size <= 192)
 		return size_index[(size - 1) / 8] - 1;
 
-	return fls(size - 1) - 1;
+	return fls(size - 1) - 2;
 }
 
 #define NUM_CACHES 11
-- 
2.29.2

Re: [PATCH bpf] bpf: Fix off-by-one error in bpf_mem_cache_idx()

[Yonghong Song]

On 1/18/23 12:46 AM, Hou Tao wrote:
> From: Hou Tao <houtao1@huawei.com>
> 
> According to the definition of sizes[NUM_CACHES], when the size passed
> to bpf_mem_cache_size() is 256, it should return 6 instead 7.

More importantly, e.g., if the size is 4096, illegal memory access may 
happen.

> 
> Fixes: 7c8199e24fa0 ("bpf: Introduce any context BPF specific memory allocator.")
> Signed-off-by: Hou Tao <houtao1@huawei.com>

Acked-by: Yonghong Song <yhs@fb.com>

> ---
>   kernel/bpf/memalloc.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/kernel/bpf/memalloc.c b/kernel/bpf/memalloc.c
> index ebcc3dd0fa19..1db156405b68 100644
> --- a/kernel/bpf/memalloc.c
> +++ b/kernel/bpf/memalloc.c
> @@ -71,7 +71,7 @@ static int bpf_mem_cache_idx(size_t size)
>   	if (size <= 192)
>   		return size_index[(size - 1) / 8] - 1;
>   
> -	return fls(size - 1) - 1;
> +	return fls(size - 1) - 2;
>   }
>   
>   #define NUM_CACHES 11

Re: [PATCH bpf] bpf: Fix off-by-one error in bpf_mem_cache_idx()

[Alexei Starovoitov]

On Wed, Jan 18, 2023 at 9:58 AM Yonghong Song <yhs@meta.com> wrote:
>
> >
> > -     return fls(size - 1) - 1;
> > +     return fls(size - 1) - 2;

Wow. Thanks.
Not sure how I missed it and why the tests didn't catch it.
test_maps goes through many key/value sizes.

Applied.

Re: [PATCH bpf] bpf: Fix off-by-one error in bpf_mem_cache_idx()

[patchwork-bot+netdevbpf]

Hello:

This patch was applied to bpf/bpf.git (master)
by Alexei Starovoitov <ast@kernel.org>:

On Wed, 18 Jan 2023 16:46:30 +0800 you wrote:
> From: Hou Tao <houtao1@huawei.com>
> 
> According to the definition of sizes[NUM_CACHES], when the size passed
> to bpf_mem_cache_size() is 256, it should return 6 instead 7.
> 
> Fixes: 7c8199e24fa0 ("bpf: Introduce any context BPF specific memory allocator.")
> Signed-off-by: Hou Tao <houtao1@huawei.com>
> 
> [...]

Here is the summary with links:
  - [bpf] bpf: Fix off-by-one error in bpf_mem_cache_idx()
    https://git.kernel.org/bpf/bpf/c/36024d023d13

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html